Download Evolution of Phenotypes

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

The Selfish Gene wikipedia , lookup

Gene expression programming wikipedia , lookup

Hologenome theory of evolution wikipedia , lookup

Sex-limited genes wikipedia , lookup

Genetics and the Origin of Species wikipedia , lookup

Evolutionary mismatch wikipedia , lookup

Sociobiology wikipedia , lookup

Mate choice wikipedia , lookup

Evolutionary landscape wikipedia , lookup

Koinophilia wikipedia , lookup

Adaptation wikipedia , lookup

Evolution wikipedia , lookup

Sexual selection wikipedia , lookup

Genetic drift wikipedia , lookup

Natural selection wikipedia , lookup

The eclipse of Darwinism wikipedia , lookup

Introduction to evolution wikipedia , lookup

Transcript
Case Studies in Evolutionary Ecology
DRAFT (notes)
Evolution of Phenotypes
[Background section about the Grant’s study of finches on the Galapagos]
[see Beak of the Finch]..
Daphne Island in the Galapagos
Four species of Galapagos finches
Peter Grant with finches
Figure 4.1 photos: http://www.lpi.usra.edu/science/treiman/heatwithin/galapagos1/galapagos1_imgs/img_0196_sm.jpg;
http://www.abc.net.au/nature/vampire/img/grantfinches.jpg
Up to this point we have been talking about the evolutionary dynamics of individual genes.
Many traits of ecological importance are determined by many genes. Those might include body
size and shape, behavior, longevity, etc. Indeed for those kinds of traits we usually have no idea
what the underlying genes are. Nevertheless those complex traits are precisely the traits that are
most often cited as the spectacular examples of natural selection and adaptation. How can we
study knowing the individual genes?
Darwin and Wallace presented their theory of evolution by natural selection before scientists
even knew that genes existed. All that is required is that traits are somehow inherited from
parents to offspring. So, even without knowing precisely why, they could look at the results
from plant and animal breeders to see how traits can be modified from one generation to the
next.
Darwin and Wallace’s theory of evolution by natural selection can be summarized by four steps:
1. There is variation in some trait of interest.
2. Individuals with traits that best match their particular environment will on average leave
more descendants than others, either because they have higher survival or higher
reproductive output.
3. Some of that variation is heritable, so offspring tend to resemble their parents.
4. The result is that the population mean will shift toward the better-adapted phenotype.
Don Stratton 2008
1
Case Studies in Evolutionary Ecology
4.1
DRAFT (notes)
A simple model of selection:
|----------------------------------------- One complete generation -------------------------------------|
Population mean at the
start of the parent
generation (before
selection)
Selection
!
Only the
largest
survive
Mean size of
the surviving
parents
Inheritance
!
Population mean at
the start of the
offspring generation
Offspring resemble the
parents that survive to
reproduce
In the case of the finches, the drought causes only some of the birds to survive to reproduction,
so the average beak size of the parents is slightly larger than the average beak size at the
beginning of the generation (because many of the birds with small beaks died). We call that
difference in size between the actual parents and the original population the “selection
differential”. Those parents then
produce offspring. Normally the
offspring will resemble their parents
somewhat, but the resemblance is not
perfect. Only to the extent that the beak
size is caused by additive effects of
alleles, will it can be inherited. Thus
there will be a scaling coefficient (that
we will call the “heritability”) that
reflects the degree to which offspring
resemble their parents.
The overall change in phenotype as a
result of selection will be a change due
to the differential survival of parents
times the degree to which that change is
inherited. That change in population
mean from one generation to the next is
called the selection "response".
4.2 Selection:
The drought of 1977 caused massive
mortality among the finches. The
Figure 4.2 Distribution of beak sizes before and after the
drought of 1977. Data from Grant 1986.
limited seed supply was soon depleted,
leaving only the large seeds that were
difficult for small-beaked birds to crack. The result was that birds with larger beaks survived at a
higher rate than birds with small beaks. The average beak depth of survivors was 9.84 mm,
compared to 9.31 mm in the general population before selection. Nevertheless some of the small
birds did survive, and some of the birds with the very largest beaks did not.
Don Stratton 2008
2
Case Studies in Evolutionary Ecology
DRAFT (notes)
Differential survival of individuals represents the action of natural selection favoring birds with
large beaks, but that is not evolution. Evolution requires a change in the average beak size from
one generation to the next. How will the increase in the number of large-beaked survivors show
up in the offspring generation? For that we also need to consider the heritability of beak depth.
4.3 Heritability
One of the other researchers on the finch team, graduate student Peter Boag, studied the
inheritance of beak depth in Galapagos finches by looking at the relationship between parent
beak depth and that of their offspring. He collected two sets of parent offspring data, once in
1976 and again in 1978. For both years he followed birds to determine which pairs belonged to
which nests. Most of the parents had been previously captured so their beak depths were
known. He then captured the offspring when they fledged and measured their beak depth. He
calculated the "midparent" beak depth (the average beak depth of the two parents) and then
compared that to beak depth of their offspring.
Here is what he found:
1976
"Midparent"
Average
beak depth of
the two parents
1978
Offspring
beak depth
Midparent
beak depth
Offspring
beak depth
8.2
8.0
8.9
8.9
8.3
7.8
9.0
9.4
8.4
8.2
9.6
9.2
9.3
7.8
9.1
9.5
8.9
8.3
9.5
9.5
8.8
8.6
9.8
9.5
9.9
8.7
10.1
9.6
10.9
8.8
9.5
10.0
9.1
8.9
9.6
9.9
9.1
9.0
9.9
9.8
9.3
9.2
10.0
10.0
9.2
9.5
10.2
10.2
9.7
9.2
10.1
10.4
9.4
9.7
10.4
10.2
9.7
9.7
10.6
10.4
9.9
9.6
10.6
10.5
10.4
9.8
10.3
10.9
10.0
9.9
10.2
10.1
10.7
10.9
Data from Boag 1983, Figure 1
Don Stratton 2008
3
Case Studies in Evolutionary Ecology
DRAFT (notes)
Figure 4.3 The heritability of beak depth in 1976.
Data from Boag 1983.
How can we use this information? The parent-offspring regression shows the way beak depth is
(on average) inherited across generations. In particular, the heritability of a trait is the slope of
thee regression line of the midparent-offspring regression. It allows us to predict how a
particular change in phenotype among the parents will be translated into a response among the
offspring. For example, if the average beak depth of the parents increases from 10.0 mm to 10.5
mm, we can use the parent-offspring regression to see how much of a change we would expect in
the offspring generation.
In 1976, the best-fit slope of offspring beak size on mid-parent beak size is 0.78. That means
that the heritability of this trait is 0.78.
Notice that the slope of this line is less than 1.0. For each unit increase in the size of the parents
there is a slightly lower gain in the offspring. Why? Because the parents beak depth is
determined in part by genetics and in part by the environment and only the genetic component of
that variation can be inherited. Therefore the heritability will always be less than 1.0.
If the phenotypic variation is determined strictly by their environment then there will be no
necessary relationship between the parent and offspring phenotype and the slope of the line
(heritability) will be 0.
That leads to a second definition of heritability: heritability is the proportion of the total
phenotypic variance that is attributable to additive genetic variance.
Optional: Repeat the analysis for the 1978 parent offspring data by graphing the offspring vs.
midparent beak sizes and finding the slope of the best-fit line.
What is the heritability for beak depth in 1978?
_________________
Are those data consistent with the heritability estimate from 1976? _________________
Don Stratton 2008
4
Case Studies in Evolutionary Ecology
DRAFT (notes)
4.4 Response to selection
Following the drought, the few surviving birds bred and the Grant's and their coworkers were
there to measure the beak size of the offspring they produced. There was still a lot of variation in
beak size, but on average, the size had increased. Beak size had evolved.
Figure 3. Evolutionary response, after the drought
4.5 The “breeders equation":
The theoretical understanding of the evolution of polygenic traits was actually first worked out
by plant and animal breeders trying to predict the yield increases for various breeding programs.
The developed the "breeders equation" that still forms the basis of our understanding of
quantitative trait evolution:
R = h2S
eq. 4.1
In words, that says that the response to selection (R) is equal to heritability (h2) times the
selection differential (S).
!
The selection differential (S) is just the difference between the mean of the population and the
mean of the individuals that reproduce. If selection is weak, then the parents who actually breed
will have a similar mean to the mean of the overall population. S will be small.
The response to selection (R) is just the difference between the mean of the parents before
selection and the mean of the offspring produced to start the next generation. Sometimes, we will
use the variable z to indicate the value of a particular trait. In that case R = "z or the change in
the trait mean across generations.
Finally the heritability (h2). is a scaling factor that translates the!strength of selection into a
realized response1. As we'll see below, the heritability can be defined by either the slope of the
regression relating midparent to offspring, or by the ratio of additive genetic variance to total
phenotypic variance.
You may wonder why the symbol h2 is used for the heritability. The practice dates back to
Sewall Wright’s original derivation, where h was a ratio of standard deviations. Therefore, the ratio of variances
(Va/Vp) must be h2. It is important to think of h2 as a single symbol, not h*h.
1
Don Stratton 2008
5
Case Studies in Evolutionary Ecology
DRAFT (notes)
The breeder's equation (eq 1) is simple, but it is very powerful because it allows a quantitative
prediction of the response to selection from a couple of easily measured parameters.
Use the breeder's equation to complete the following table:
What is the selection differential? What is the predicted response? How does that compare to the
observed response?
Average Beak Depth at start of 1976 (before selection) 9.31 mm
Beak Depth of survivors, who lived to breed in 1978 9.84 mm
Selection Differential
Heritability 0.78
Predicted response to selection
Predicted mean of the offspring born 1978
Average Beak Depth of the offspring born in 1978 9.70 mm
4.6
More about the genetic basis of quantitative traits
Natural selection acts on the variation that is present in the population. The total phenotypic
variance can be partitioned into a genetic and an environmental component. In turn, the genetic
variance is composed an additive genetic component (caused by the average effects of alleles)
and a non-additive or dominance component, caused by particular combinations of alleles.
Total Phenotypic Variance (Vp)
Genetic Variance
Additive Component (VA)
Environmental Variance
Dominance Component (VD)
VE
Parents only pass a single allele on to their offspring, so only the additive or average effects of
the alleles can be inherited. Effects due to particular combinations of alleles (dominance) are
broken up each generation. Therefore the evolutionary response to selection will depend upon
the magnitude of the additive component of genetic variance, symbolized VA.
The ratio of additive genetic variance to total phenotypic variance is another way to define the
heritability2.
V
h2 = A
eq 4.2
VP
2
Although we won't derive it, this definition is mathematically equivalent to the offspring-midparent regression.
!
Don Stratton 2008
6
Case Studies in Evolutionary Ecology
DRAFT (notes)
If we wanted, we could also partition the environmental component into effects of the rearing
environment, developmental defects, nutritional status, etc. etc.
4.7 More about selection
To see the pattern of survival more clearly, it helps to combine the data into larger size categories
so some of the random variation is averaged out. What proportion of each size class survived the
drought?
Size class
N before
selection
N
Survivors
Less than 7
21
1
7.0 to 7.9
34
2
8.0 to 8.9
186
10
9.0 to 9.9
294
25
10.0 to 10.9
186
43
11.0 to 11.9
25
9
Greater than 12
5
0
Survival rate
Use those data to graph fitness (survival) vs. beak size.
How do you explain the low survival of the birds with the very largest beaks?
Is it just chance that none of those 5 survived? How could you tell?
Don Stratton 2008
7
Case Studies in Evolutionary Ecology
DRAFT (notes)
This graph is called the "fitness function". Fitness functions describe the way fitness relates to a
particular trait, in this case the relationship between survival during drought and beak depth.
Fitness functions with positive slopes mean that selection favors an increase in the trait. Fitness
functions with negative slopes mean that selection favors a decrease in the trait. Curved fitness
functions indicate stabilizing selection of disruptive selection (Fig 4.4).
Figure 4.4. Fitness functions and their effects on phenotypes. The solid line shows the frequency
distribution before selection and the dashed line is the distribution after selection.
Directional selection causes a
change in the mean value of a
trait.
4.8
Stabilizing selection causes a
decrease in the variance of the
trait, with no change in mean.
Disruptive selection causes an
increase in the variance of the
trait.
More about heritability.
When only one of the two parents is measured, the slope of the regression estimates only
1 2
h .
2
Therefore you need to double the slope of the mother-offspring or father-offspring regression to
obtain the heritability.
For example, Boag compared his midparent-offspring regression to the father-offspring
and
!
mother-offspring regressions. Notice how the regressions on single parents both have much
lower slope than the midparent regression. What is the estimate of heritability for each of those
regressions?
Bill depth
Regression
slope
MidparentOffspring
Father-Offspring
Mother Offspring
0.72
0.22
0.39
Heritability
Don Stratton 2008
8
Case Studies in Evolutionary Ecology
DRAFT (notes)
Why do you think the estimate of heritability using mothers is so much higher
than the estimate using the father-offspring regression?
One of the assumptions of using parent-offspring regressions to estimate the heritability is that
the only source of resemblance is shared genes. In many organisms it is common to see
"maternal effects". For example, well fed, females may lay large eggs, which produce large
offspring. That produces a resemblance between mothers and offspring that is strictly
environmental. That kind of environmental correlation is less common through males, who
generally provide only sperm.
Second, the fathers at the nest aren't always the birds that actually sired the offspring. In G.
fortis approximately 20% of the offspring are produced through extra-pair matings. DNA
samples revealed that 44 of 223 offspring did not match their putative father's genotype and must
have been sired by another male. If the non-genetic fathers are excluded from the fatheroffpsring regression, then the slope increases to b=0.36.
4.8.1 Some common misconceptions about heritability:
• Heritability of zero does not mean the trait is not determined by genes. It means that the
variation in the trait does not have a genetic basis. Some traits that are universal (e.g.
presence of two eyes in humans) are still determined by our genes, but there is no
variation in that trait so the heritability would be zero.
• Environmental effects can still cause differences between populations, even if the
heritability is not zero.
• Two populations with the same genetic variance may have very different heritabilities if
one is subject to more environmental variance than the other.
Don Stratton 2008
9
Case Studies in Evolutionary Ecology
4.9
DRAFT (notes)
EXTRA: The relationship between selection on single loci and the breeders
equation:
1 gene
To understand the genetic basis of quantitative traits, it is important
to think about the effect of a particular allele, not simply its
presence or absence. A single locus can produce three discrete
phenotypes, but as more and more loci contribute to a trait the
phenotypic distribution comes closer and closer to a normal (bell
shaped) distribution.
Sum of 2 genes
Sum of 4 genes
Any normal distribution can be described by the mean and
variance. The breeder's equation uses those aggregate measures of
the trait to predict the response to selection without explicitly
following individual alleles. Nevertheless it is important to realize
that at each of the underlying loci selection will change allele
frequencies following exactly the same model that we used for
other Mendelian loci.
Let's imagine that we are studying the size of some organism. To
keep things simple, we will assume that size is controlled by a
large number of independent genes, each with equal effect. We'll
assume that there are two alleles, "big" and "small". The "big"
allele at each locus adds 1/2 a unit of size and the "small" allele
subtracts 1/2 unit. Therefore the net effect on size for each locus
will be the sum of the values of the two alleles: the net effect will
be -1 if both copies are the "small" allele, 0 if it is heterozygous,
and +1 for the other homozygote where there are two doses of the
"big" allele. Ultimately, the trait (in this case size) will be
determined by the total dosage of "big" alleles summed over all
loci.
At each locus the additive genetic variance will be (assuming no dominance)
VA = 2 pqa2
eq. 4.3
What this equation makes clear is that the additive genetic variance depends on the allele
!
frequencies (p and q) as well as the squared value
of the alleles (a2). If one of the alleles goes to
fixation (p=0 or q=0) then there will be no genetic variation at this locus.
Why are we only concerned with the additive component of genetic variance?
Example: Imagine that a particular locus has two alleles B and b. Each copy of the B allele
increases the size of an organism by 2 mm (a=2) and there is no dominance. Large individuals
have a fitness advantage, so allele B gradually increases to fixation. What happens to the genetic
variance as B increases in frequency?
Don Stratton 2008
10
Case Studies in Evolutionary Ecology
DRAFT (notes)
Frequency of
B allele
VA
0.50
2.0
0.75
0.95
1.00
This represents a general phenomenon that directional selection acts to decrease the genetic
variance for a trait. After many generations of selection, if there is no new source of genetic
variation by mutation, the additive variance will decline to zero and there will no longer be any
response to selection.
A corollary to this is that for traits that under strong selection, there should be little genetic
variation remaining in the population.
p=0.50
p=0.75
p=0.95
(parameters: a=2, d=0, Ve=0.5)
As allele B goes to fixation, the genetic variation contributed by this locus (dashed line) declines. Note that there is
some variation in the phenotype of each of the three genotypes, because of environmental variation.
Don Stratton 2008
11